ritual-meta-non-interactive-bias

To stop constant agent prompts, apply a non-interactive bias that enforces a search-before-asking policy, forcing the agent to resolve tasks through internal discovery and contextual reasoning before requesting user input.

Budget-aware reasoning tracks turn-based budgets and adjusts agent behavior based on remaining capacity, ensuring autonomous resolution stays within defined limits while only prompting the user when a budget overrun occurs.

An autonomous agent should stop and ask the user only when hitting clear exit conditions: needing credentials, facing a correctness fork, or encountering a budget overrun, otherwise proceeding autonomously.

A search-before-asking policy is a resolution hierarchy where agents prioritize self-discovery, contract inspection, and contextual reasoning to automate decisions and announce them without requesting user permission.

You can enforce autonomous decision-making by applying a skill that manages the resolution hierarchy and turn-based budget tracking, announcing standard decisions automatically without requiring custom control flow.

The limitation of non-interactive bias is that it cannot handle tasks requiring external credentials or correctness forks autonomously, and it will force a user prompt if the workflow exceeds the allocated turn budget.